cubecl-matmul 0.8.1

CubeCL Matrix Multiplication Kernels Engine
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

#[macro_export]
macro_rules! testgen_matmul_mma_tile {
    ($algorithm: ty, $precision: ty, $tiling_scheme_builder: expr) => {
        use $crate::components::TileSize;

        // AMD f16/bf16
        #[cfg(not(target_os = "macos"))]
        mod t16x16x16 {
            use super::*;

            $crate::testgen_matmul_mma_partition!(
                $algorithm,
                $precision,
                $tiling_scheme_builder.with_tile_size(TileSize {
                    m: 16,
                    n: 16,
                    k: 16
                })
            );
        }

        // Nvidia f16/bf16
        #[cfg(not(target_os = "macos"))]
        mod t16x8x16 {
            use super::*;

            $crate::testgen_matmul_mma_partition!(
                $algorithm,
                $precision,
                $tiling_scheme_builder.with_tile_size(TileSize { m: 16, n: 8, k: 16 })
            );
        }

        // Nvidia tf32
        #[cfg(not(target_os = "macos"))]
        mod t16x8x8 {
            use super::*;

            $crate::testgen_matmul_mma_partition!(
                $algorithm,
                $precision,
                $tiling_scheme_builder.with_tile_size(TileSize { m: 16, n: 8, k: 8 })
            );
        }
    };
}